Hello and welcome to the Stone News, a newsletter where we discuss every three months the most recent and relevant studies in stone disease.

Suscribe now

Dear Stone Fans. Stone disease affects 10–15% of the global population; the quest for better prevention and management strategies must continue to develop, as many people suffer from this condition. 

This month's selections explore three facets of modern stone care: the chemistry of urinary pH manipulation, the mystery of why some forgotten stents calcify while others remain pristine, and how artificial intelligence is stepping into the operating room to predict who might land in the ICU.

From citrate to machine learning, these studies remind us that stone disease sits at the fascinating intersection of metabolism, materials science, and now, computer science. Please enjoy this spring’s Stone News.

1'

This review examines the role of urine pH in kidney stones. Urine pH critically influences the formation of various stone types: low pH (<5.5) promotes uric acid stones, while high pH favors calcium phosphate and struvite stones; the relationship with calcium oxalate stones remains complex and under investigation.

Regarding pharmacological management, potassium citrate is the cornerstone alkalizing agent, achieving complete uric acid stone dissolution in up to 70% of cases and nearly 90% stone-free rates for ureteral uric acid stones by raising urine pH to 6.0–6.5, where uric acid converts to its more soluble urate form. Combining citrate with theobromine (Lit-Control^® pH Up) may improve efficacy and tolerability of the treatment.

For struvite and calcium phosphate stones that thrive in alkaline urine, acidifying agents such as L-methionine (Lit-Control^® pH Down, reducing struvite supersaturation by 34%) and ammonium chloride effectively lower urine pH. 

For calcium oxalate stones, phytate (Lit-Control^® pH Balance) emerges as a promising pH-neutral alternative, demonstrating efficacy equivalent to citrate at significantly lower doses (1.5 mg vs. 800 mg) without the risk of calcium phosphate precipitation. The authors conclude that while effective pH modification strategies exist for specific stone types, standardization of monitoring methods and further research are needed to optimize individualized treatment approaches.

1'

This retrospective study examined predictors of encrustation in 85 patients with forgotten ureteral stents (mean dwell time 14.6 months), finding that not all forgotten stents become encrusted — 54 of 85 patients developed encrustation despite prolonged indwelling. 

The strongest independent predictors were low preoperative stone Hounsfield Unit values (<647.5 HU; p=0.003) and increased ureteral diameter (>6.9 mm; p=0.002)—the authors hypothesize that low-HU stones (e.g., uric acid) have higher recurrence rates, promoting encrustation around the foreign body.

The findings suggest that patients receiving stents after low-density stone surgery may benefit from closer follow-up, earlier removal, and from pharmacological modification of urinary pH. The study also validated three encrustation classification systems, with the KUB scoring system demonstrating superior correlation and predictive utility for surgical planning compared to V-GUES and FECal systems.

1'

This study developed machine learning (ML) models to predict postoperative sepsis following endourologic kidney stone surgery (ureteroscopy, SWL, or PCNL), addressing a significant gap in urosepsis risk stratification that traditionally relies on clinical judgment alone. Among 382 patients from a tertiary center (2020–2023), five ML algorithms were compared, with the random forest model achieving the best performance (91% accuracy, AUCROC 0.88, Brier score 0.09).

Feature importance analysis identified the top five predictors as preoperative hemoglobin, HbA1c, stone size, length of surgery, and BMI—notably, preoperative anemia emerged as the strongest predictor, a novel finding in endourology literature potentially explained by immune dysfunction in anemic patients.

Regarding AI's role in clinical management, the model offers several practical applications: it can guide preoperative optimization by identifying high-risk patients who may benefit from anemia correction or improved glycemic control before surgery; inform perioperative decision-making such as broader-spectrum antibiotics, prolonged recovery observation, or increased fluid resuscitation; and enhance postoperative monitoring intensity for flagged patients. The authors deployed the model as a free online calculator (https://urol.shinyapps.io/sepsis_predict/) for clinical use and further validation, demonstrating how AI tools can be translated into accessible point-of-care applications.